22 May 2014 Near-infrared spectroscopy and pattern-recognition processing for classifying wines of two Italian provinces
Author Affiliations +
This paper presents an experiment making use of the near-infrared spectrum for distinguishing the wines produced in two close provinces of Abruzzo region of Italy. A collection of 32 wines was considered, 18 of which were produced in the province of Chieti, while the other 14 were from the province of Teramo. A conventional dual-beam spectrophotometer was used for absorption measurements in the 1300-1900 nm spectroscopic range. Principal Component Analysis was used for explorative analysis. Score maps in the PC1-PC2 or PC2-PC3 spaces were obtained, which successfully grouped the wine samples in two distinct clusters, corresponding to Chieti and Teramo provinces, respectively. A modelling of dual-band spectroscopy was also proposed, making use of two LEDs for illumination and a PIN detector instead of the spectrometer. These data were processed using Linear Discriminant Analysis which demonstrated satisfactory classification results.
© (2014) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
A. G. Mignani, A. G. Mignani, L. Ciaccheri, L. Ciaccheri, B. Gordillo, B. Gordillo, Andrea A. Mencaglia, Andrea A. Mencaglia, M. L. González-Miret, M. L. González-Miret, F. J. Heredia, F. J. Heredia, A. Cichelli, A. Cichelli, "Near-infrared spectroscopy and pattern-recognition processing for classifying wines of two Italian provinces", Proc. SPIE 9106, Advanced Environmental, Chemical, and Biological Sensing Technologies XI, 91060G (22 May 2014); doi: 10.1117/12.2051914; https://doi.org/10.1117/12.2051914

Back to Top